Molecular Weight Determination of Sendai RNA by Dimethyl Sulfoxide Gradient Sedimentation

Copyrighti 1973 American SocietyforMicrobiology Printed inU.S.A.

Molecular

Weight

Determination

of

Sendai

RNA

by

Dimethyl

Sulfoxide

Gradient

Sedimentation

DANIEL KOLAKOFSKY AND ANDRtE BRUSCHI

Department of Molecular Biology, UniversityofGeneva, Geneva, Switzerland

Received for publication 14 December 1972

The molecularweight of the largeRNA of Sendai virus has beendetermined by sedimentation analysis in sucrose gradients containing 99% dimethyl

sulfoxide

(DMSO)

tobe 2.3 x 106.Sendai RNA recovered from 99% DMSOwas

found tocosediment with nondenatured Sendai RNA at 46 to 48s in ordinary sucrose gradients. The molecular weight value of 2.3 x 106 is considerably

smaller than the estimates of 6 x 106to7 x 106 determined under nondenatur-ingconditions, suggesting auniquestructurefor Sendai RNA.

Sendai virions contain

a

large

single-stranded RNA

(3,

4,

18)

which is

encapsulated

in a

nucleocapsid

structure

morphologically

similar

to

tobacco mosaic virus

(TMV)

but

approximately three times

as

long

(14,

15). The

molecular

weight

of

parainfluenza viral RNA

(Sendai, Newcastle disease virus

[NDV],

sim-ian virus

5

[SV5])

has

been estimated

by

sucrose

gradients under

nondenaturing

condi-tions to

be

approximately

6 x 106 to 7 x 106

(3,

9,

11). We have

recently measured the size of

this RNA

under

completely

denaturing

condi-tions,

99%

dimethyl

sulfoxide

(DMSO) (28),

and have found the molecular weight

to

be

2.3

x

106,

a

value

considerably

smaller than

those

obtained from sedimentation

analyses under

nondenaturing conditions.

Since

Sendai RNA,

which

was

recovered

from a

DMSO

gradient

by

ethanol precipitation,

was

found

to

cosediment

in

ordinary

sucrose

gradients with

non-denatured Sendai RNA,

it is

unlikely

that this

lower-molecular-weight determination

wasdue

to

fragmentation

of

the

Sendai

genome (see

Discussion). The implications

of this

difference

are

discussed.

MATERIALS AND METHODS

Viral RNA.

3"P-Sendai

virus

(Harris

strain,

ob-tained fromR. D. Barry,Cambridge, England) and

32P-NDV(Californiastrain, obtained from B.Colby,

Storrs, Conn.) were grown in 9-day-old embryonated

chickeneggs labeled with 0.5mCiof 32p,per egg (at

timeofinfection) for 72 h at 34 C and 48 h at 39 C,

respectively. -HSendai virus was labeled by

prein-jectingthe eggs one day before infection with 0.3 mCi

of 3H-uridine (20 Ci/mmol). Virus was purified as

previously described (19), except that the viruswas

bandedonce in 25to65%sucrosegradients (4 h at

26,000 rpm,7C, in theSpinco SW27rotor) instead of

banding intartrategradients.

RNAwasisolated from the virionsasfollows.Virus

suspensions, at approximately 5 mg/ml in TNE

buffer (19), were made 0.2% in sodium dodecyl

sulfate(SDS)and 100Mg/mlin ProteinaseK(Merck)

and incubated for 20 min at25C. The solutionwas

then twice extracted withanequalvolume ofphenol

(saturated with TNE) at25C, and thewaterphase

was made 0.2 M inNaOAc, pH 5.3. The RNAwas

precipitated with two volumes of ethanol and left

overnight at -20C.The RNAwasthen collectedby

centrifugation, dissolvedindistilledwater, and

cen-trifuged in sucrose gradients as described below.

Fractionscontaining the fastsedimentingRNApeak

(cf.Fig.1A)werepooled, andthe RNAwasrecovered

from the gradient by ethanol precipitation. S2p_

labeled Sendai and NDV RNA, on isolation,

con-tained approximately20,000 countsper min per

jg,

'H-labeledSendaiRNA,3,300countsper min per

jg.

Other RNAs. Radioactive cellular RNAs were

obtained by labeling nonconfluent uninfected

cul-turesofmousekidney cells with4

MCi

of"C-uridine

(60mCi/mmol) permlorchickenembryofibroblast

cultures with 25uCi of 3H-uridine (20Ci/mmol) per

ml for 27 h. RNA was extracted as described by

Achesonetal.(1). `C-RNA contained 10,000 counts

perminper

gg,

and3H-RNA, 150,000 countsper min

per

Ag.

'H-labeled 45s rRNA precursor (220 counts per

min perMg), the kind gift of M.-E. Mirault and K.

Scherrer,wasisolated from Hela nucleoli labeled for

15 min with methionine-methyl-3H CT,, as

previ-ouslydescribed (22).

Sucrosegradient sedimentation. Linear sucrose

gradients (5-23%) contained 0.1 M LiCl, 10 mM Tris-hydrochloride (pH 7.4), 4 mM EDTA, and 0.1%

615

on November 10, 2019 by guest

http://jvi.asm.org/

KOLAKOFSKY ANDBRUSCHI

SDS. LiCl was used in place of NaCl to aid the

solubility of SDS at lower temperatures.

Centrifuga-tions were carried out for 105 min at 50,000 rpm, at 7

C, in the Spinco SW56 rotor. Gradients were

frac-tionatedbypuncturing thebottom of the tube,and

drops were collected on paper discs. The paper discs

were then dried, washed first in 6% trichloroacetic

acid and then in ethanol, dried, and counted by liquid scintillation.

DMSO gradient sedimentation. DMSO

gradi-ents were prepared, and sample application was

carried out as described by Acheson et al. (1).

Because the absolutesedimentation rate of RNA was

found to varyconsiderably in different preparations

of the DMSO solutions (presumably due to the

hygroscopic nature of the solvent),DMSOsolutions

were prepared in batchquantities, divided into equal

parts, and kept sealed in a desiccator to ensure

reproducibility from runto run. Gradients reported

here were centrifugedfrom 3 to 4.5 h at 48,000 rpm,

26 C, in the Spinco SW56 rotor. Gradients were

fractionated and countedas described above.

RESULTS

In sucrose gradients

containing 0.1 M NaCl,

the sedimentation

value of Sendai RNA has

46s

SUCF

6-3

/

EI 10 20

32S

8

^

DM

c

W

been

measured

as 57s

relative

to

TMV RNA

(3), and 50s relative

to28s

and

18s

rRNA

(18).

The RNA of

two

other

type II

myxoviruses,

NDV and SV5,

have both been measured

as

50s

(9, 16).

Under the conditions

we

have used

(0.1

M

LiCl,

see

Materials and

Methods), the large

RNA of four different

preparations of

Sendai

virus

was

found

to

sediment

asa

sharp peak

at

46 to 48s

relative

to 28s

and 1&s rRNA

(Fig. 1A). However,

when the

same RNA was run in

gradients containing

99%

DMSO, Sendai RNA

was

found

to

sediment just ahead

of

28s

rRNA,

at 32s.

(All

sedimentation values

in

DMSO

reported

here

are

relative

to the

ribosomal

RNAs

taken

as 18s

and

28s, in

DMSO).

Be-cause

of

this unexpected finding, the

DMSO

gradients

were

calibrated with

an

RNA marker

expected

to

sediment faster than

Sendai

RNA

in

DMSO,

namely,

the 45s ribosomal

RNA

precursor (29).

As

can

be

seen in Fig. 2, the 45s

rRNA

precursor (46s under our conditions on

sucrose) was found to sediment at 43s in

DMSO. The molecular weight of Sendai RNA

using the

abovementioned markers (and R17

)SE

A

10

30 40 0

E

B

20

fraction number

FIG. 1. Comparisonof32P-Sendai RNA and14C-cellularRNA

(4s,

18s,and28sRNA)on

ordinary

sucrose

(A)andDMSOgradients

(B).

Gradient

(A)

contained0.6

sg of

32P-Sendai RNA

(3,200

counts/min)and 0.5yg

of"IC-cellular RNA. Gradient

(B)

contained 0.6yg

of

82P-SendaiRNA and 2.0ggof "4C-cellular RNA.

Con-ditionsofcentrifugationand

counting

aredescribed in Materials and Methods.

616

J.VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:2.493.121.395.336.621.2]

VOL. 11, 1973

MOLECULAR WEIGHT

RNA

xe

E

I

E

2-2

10 20 30 40

[image:3.493.47.232.60.399.2]

fraction number

FIG. 2. Comparison of 'H-45s rRNA precursor,

"4C-cellular

RNA,

and32P-Sendai RNAon

ordinary

sucrose (A) and DMSO

gradients (B)

and

(C).

Gradient (A) contained2.9

Mg of

3H-45s rRNA and 2.0

Mg

of "4C-cellularRNA. Gradient

(B)

contained

2.9 ug of 3H-45s rRNA and 2.0

Mg of

"4C-cellular

RNA. Gradient(C) contained 2.9

Mg of

'H-45srRNA

and 2.0Mg of32P-Sendai RNA (4,100 counts/min).

RNA, data

not

shown)

was

estimated

to

be

2.3

x

106

(Fig. 3).

The

possibility that,

as

for

Rous

sarcoma

virus

(RSV)

RNA

(10, 12,

23),

the

large

differ-ence

in

sedimentation

rate

of

Sendai RNA

in

DMSO and sucrose

gradients

might be due to

disaggregation

of

a

segmented

genome

in

DMSO

was

next

examined.

"P-labeled Sendai

RNA

was

prepared

by

first

sedimenting

it on

a

DMSO gradient, and the RNA sedimenting at

32SDMSO

was recovered from the

gradient

by

ethanol

precipitation after

the

profile of the

gradient

had

been determined by

Cerenkov

counting (cf. Fig. 1B). This Sendai RNA was

then

mixed with

nondenatured,

8H-labeled

Sendai

RNA

and cell

RNA

markers and

resedi-mented

in

an

ordinary sucrose gradient. The

results, shown

in

Fig. 4,

demonstrate that

Sendai

RNA

sediments

at the

same

position, 46

to

48s, even

after recovery from the DMSO

gradient.

Thus, unlike RSV RNA, the

sedimen-tation

behavior of

Sendai.

RNA in sucrose

gradients is unchanged

by DMSO treatment,

and 46 to 48s

Sendai RNA appears to be a

single, continuous RNA chain of

2.3

x

101

(see

Discussion).

Although the molecular weight of Sendai

RNA has

not previously

been measured under

denaturing

conditions, Duesberg

(10) has

measured the molecular

weight of the

large

RNA of

NDV, another type II

myxovirus, in

DMSO gradients as 6.3 x 10' daltons. We have

therefore

compared

the

large

RNAs of Sendai

and NDV on both

ordinary

sucrose

and

DMSO

gradients. We find that Sendai and NDV RNA

are

indistinguishable

in

sedimentation

rate

in

both

ordinary

sucrose

(reference

4,

Fig. 5A)

and

DMSO

gradients

(Fig.

5B),

and therefore

con-clude that the

large

RNA of

NDV,

like that of

Sendai

virus,

has

a

molecular

weight

of

2.3

x

106.

DISCUSSION

As

previously

mentioned, Duesberg (10)

has

reported the molecular

weight

of

NDV RNA in

4

3

I

0

x

3:

0

2

2

20

30 40

50

Relative

SDMSO

FIG. 3. Determination of the molecular weight of

Sendai RNA bysedimentation inDMSOgradients.

RNAs used as reference markers were: 18s rRNA,

0.65 x 10' (20); R17RNA, 1.05 x 10' (12); 28s rRNA,

1.65 x

106

(20); 45s rRNA precursor, 4.1 x

106

(27).

All

SDS,O

arerelative to 28s and 18s rRNA taken as

28& and 18s in DMSO. The molecular weight of

Sendai RNA, represented by an open circle, is

determined as 2.3 x 10 .

617

I

on November 10, 2019 by guest

http://jvi.asm.org/

[image:3.493.248.436.357.589.2]

KOLAKOFSKY AND BRUSCHI

2

0

10 20 30

I0

C-)

5

E

6.)

0

fraction

number

FIG. 4. Comparison of DMSO-denatured and nondenaturedSendai RNAon sucrosegradients. 32P-Sendai

RNA (3.5,g,10,400counts/min)wasfirst sedimentedinaDMSO gradient (cf. Fig.1B),and theprofileof the

gradientwasdetermined by Cerenkovcounting. RNApresentinthe threepeak tubeswasrecoveredfrom the

gradient by ethanol precipitation and dissolved in distilled water. A sample of this material (1,200

counts/min)wasmixedwith 1.1 Mgof nondenatured sH-Sendai RNA and0.09ugof 3H-cellular RNA andthen

sedimented inanordinarysucrosegradient.

DMSO gradients to be 6.3 x 106, whereas we

find both NDV and Sendai RNA to be 2.3 x

106, using similar methods. The molecular weight value of 6.3 x 106l, however, was

ob-tained by means of the empirical equation

determined by Strauss et al. (28) for the absolute sedimentation rate of RNA in the analytical ultracentrifuge and not by

extrapo-lation fromthesedimentation rateofRNAsof known molecularweight determined under si-milar conditions, i.e., sucrose gradients in DMSO. Because of the large effects ofsucrose on theviscosity of solutions, it isunlikely that

the empirical equation determined by Strauss

etal. wasapplicablein this case.

The large difference in molecular weight between thatexpectedfora46to48sRNA(27) and that measured under denaturing

condi-tions can be accounted for if Sendai RNA

contained a great deal more secondary

struc-ture than the ribosomal RNAs. Sendai RNA

has recently been shown to self-anneal, to

variableextents(16-60%,references24,25)ina

concentration-dependent fashion. However, as

Portner and Kingsbury (24) havepointed out,

the fact that the RNA had been degraded to

approximately 10s during the annealing

reac-tiondidnot allow themtodistinguishwhether

thisannealingwasintra-orintermolecular.We

have also found that Sendai RNAself-anneals

(15-25%, datanotshown),and if thisannealing

is in fact intramolecular, it could explain the

abovementioned discrepancy in sedimentation

rateunderdenaturing and nondenaturing

con-ditions.

There is, however, another possibility to

explain this discrepancy. Ourconclusions that the 46to48sRNA iscomposed ofasingle chain

of 2.3 x 106 molecular weight is basedon the

unaltered sedimentation rate of Sendai RNA

on sucrose gradients after

DMSO

treatment

(cf. Fig. 4). However, if Sendai RNA were

capable of rapid andquantitative reassociation (e.g., byannealing ofcomplementary ends) due

tothe increase inconcentration when theRNA is removed from the denaturing solvent by ethanol precipitation and subsequent solution

in water, then the 46 to 48s RNA could be

composed ofmorethanoneRNAchain of2.3 x 106 molecular weight. Although this latter possibility seems unlikely, there is some

evi-dence in its favor. In thevirion, parainfluenza virus RNAoccursintheformofanucleocapsid

which is remarkably similar in structure to

TMV(6), but

approximately

3.3 timesaslong

(6-8, 14, 15, 17). These nucleocapsids have

beenshowntocontainfrom3.7to4.1% RNA(7, 14), only slightly less than the 5.1% RNA

determined for TMV (19), whose RNA has a

molecular weight of 2.1 x 106 (5). These

nucleocapsids should therefore contain 5.0 x

106 to 5.6 x 106 daltons of RNA. Since the

molecular weight of Sendai RNA as

deter-mined by DMSO gradient sedimentation is 0~

L)

C.) x

E

618 J. VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:4.493.111.400.60.273.2]

I0

I4

N

x

C.

Ci4

C,)

6

3

N

I

°

x

cL u

I

10

5

O

I

10

20

30

40

[image:5.493.101.382.64.411.2]

frraction number

FIG. 5. Comparison of32P-NDV RNA and

3H-Sendai

RNA on

ordinary

sucrose

(A)

and

DMSO

gradients

(B).Gradient

(A)

contained38,000 counts/min

of

32P-NDV virions and 0.8

Mtg of

3H-Sendai RNA. Thesamples

weresuspended in 75

Mliters

ofwater,made1%inSDS,and thensedimentedon an

ordinary

sucrose

gradient.

Gradient(B) contained0.4

Mgg

of32P-NDV RNA (4,600

counts/min)

and1.6

Mg of

'H-SendaiRNA.

well

below this

value, Sendai

nucleocapsids

should therefore contain

more

than

one

chain

of

RNA. It is

possible

that the

structure

respon-sible

for

these

multiple

chains

of

RNA in the

nucleocapsid

is

reassumed

on

removal

of

the

RNA from

DMSO;

this reassociation would

of

course

be concentration

dependent.

In

conclusion,

we

would like

to

add

a noteof caution.

This

discrepancy

between

sedimenta-tion

value

in

ordinary

sucrose

and

DMSO

gradients

is not

unique

to

Sendai RNA.

Arif

and Faulkner(2) have recently shown that 42s

Sindbis

RNA

sediments

only

slightly

faster

than

28s rRNA

in

DMSO

gradients, and that

this

RNA, when

recovered from a DMSO

gradient, coelectrophoreses

on

polyacrylamide

gels

with untreated 42s RNA. It seems clear that these animal viral RNAsare different from

rRNAand

bacteriophage

RNA in this respect.

Even

though these RNAs have been purified in

a

similar

manner as

rRNA and

bacteriophage

RNA,

the

possibility

cannot

be excluded that

they

still contain material which

can

differen-tially

affect

their

buoyant density

inwater

and

DMSO (the

inability of Simmons and Strauss

[26]

to

sediment

Sindbis viral RNA

as a

discrete band

in

DMSO

gradients

is

note-worthy

in

this respect), thus

making molecular

weight estimation of these viral RNAs

in

DMSO

gradients unreliable using rRNA and

bacteriophage RNA

as

markers.

ACKNOWLEDGMENTS

We thank Peter Bromley for helpful discussion, and

Pierre-FrangoisSpahr for generous advice, encouragement, and support.

Thiswork wassupported by research grant no. 3.816.72 from the Fonds National Suisse de la Recherche Scien-tifique.

LITERATURE CITED

1. Acheson, N. H., E. Buetti, K. Scherrer, and R. Weil.

VOL.11,

619

on November 10, 2019 by guest

http://jvi.asm.org/

KOLAKOFSKYANDBRUSCHI

1971. Transcription of the polyoma virus genome:

synthesis andcleavage ofgiant latepolyomaspecific RNA. Proc. Nat. Acad.Sci. U.S.A.68:2231-2235. 2. Arif, B. M., and P. Faulkner. 1972.Genome of Sindbis

virus.J. Virol.9:102-109.

3. Barry, R. D., andA.G.Bukrinskaya.1968. The nucleic acid of Sendai virus and ribonucleic acid synthesis in cells infectedby Sendaivirus. J. Gen. Virol.2:71-79. 4. Blair, C. D., and W. S.Robinson. 1968.Replicationof Sendai virus, comparison of viral RNA and virus specific RNA synthesiswith Newcastle disease virus. Virology35:537-549.

5. Boedtker, H. 1960.Configurational propertiesoftobacco mosaicvirusribonucleic acid.J. Mol. Biol.2:171-188. 6. Choppin, P. W., and W. Stoeckenius. 1964. The

mor-phology of SV5 virus. Virology 23:195-202.

7. Compans, R. W., and P. W. Choppin. 1967. Isolation and properties of the helical nucleocapsid of the parainfluenzavirusSV5.Proc. Nat. Acad. Sci.U.S.A. 57:949-956.

8. Compans, R.W.,and P. W. Choppin, 1967. Thelength of the helicalnucleocapsidofNewcastle disease virus. Virology33:344-346.

9. Compans,R.W.,and P. W.Choppin.1968. Thenucleic acid ofparainfluenzavirusSV5.Virology35:289-296. 10. Duesberg, P. H. 1968. Physical properties of Rous

sarcoma virus RNA. Proc. Nat. Acad. Sci. X.S.A. 60:1511-1518.

11. Duesberg, P. H., and W. S. Robinson.1965.Isolation of the nucleic acid of Newcastle disease virus (NDV). Proc. Nat. Acad. Sci. U.S.A. 54: 794-800.

12. Erikson, R. L. 1969. Studies onthe RNA from avian myeloblastosisvirus.Virology 37:124-131.

13. Gesteland,R.F.,and H.Boedtker.1964.Some physical properties ofbacteriophage R17 and its ribonucleic acid.J. Mol. Biol.8:496-507.

14. Hosaka, Y. 1968. Isolation and structure of the

nu-cleocapsidof HVJ.Virology35:445-457.

15. Hosaka, Y.,H.Kitano,and S.Ikeguchi.1966. Studieson

the pleomorphism of HVJ virions. Virology 29:205-221.

16. Kingsbury,D. W. 1966. Newcastle disease virus RNA.I.

Isolation and preliminary characterization ofRNA fromvirusparticles.J. Mol. Biol. 18:195-203. 17. Kingsbury, D. W., andR. W. Darlington.1968.Isolation

and propertiesofNewcastle diseasevirus nucleocap-sid. J. Virol.2:248-255.

18. Kingsbury, D. W., A. Portner, and R. W. Darlington. 1970. Properties of incomplete Sendai virions and subgenomic viral RNAs. Virology42:857-871. 19. Knight, C. A., and B. R. Woody. 1958. Phosphorous

content of tobacco mosaic virus. Arch. Biochem. Biophys.78:460-467.

20. Kolakofsky, D. 1972. Transfer ribonucleic acid

nu-cleotidyltransferase andtransfer ribonucleic acid in Sendai virions. J. Virol. 10:555-559.

21. McConkey, E. H., and J. W. Hopkins.1969. Molecular weightsofsomeHelaribosomalRNAs. J. Mol. Biol. 39:545-550.

22. Mirault, M.-E., and K. Scherrer. 1971. Isolation of preribosomes fromHela cellsand their characteriza-tion by electrophoresis onuniform and exponential-gradient polyacrylamide gels. Eur. J. Biochem. 23:372-386.

23. Montagnier, L.,A.Golde,and P.Vigier.1969. Apossible subunit structure of Rous sarcoma virus RNA. J. Gen.Virol. 4:449-452.

24. Portner, A., and D. W. Kingsbury. 1970. Complemen-tary RNAs in paramyxovirions and paramyxovirus-infected cells. Nature(London)228:1196-1197. 25. Robinson, W. S. 1970. Self-annealing of subgroup 2

myxovirusRNAs.Nature(London) 225:944-945. 26. Simmons,D.T.,and J.H.Strauss. 1972.Replicationof

Sindbisvirus. I. Relative size andgeneticcontent of

26Sand 49SRNA. J. Mol. Biol.71:599-613. 27. Spirin, A. S. 1963. Some problems concerning the

macromolecularstructureof ribonucleic acids.Progr. Nucleic Acid Res. 1:301-345.

28. Strauss, J. H. Jr., R. B. Kelly, and R. L. Sinsheimer. 1968. Denaturation of RNA withdimethyl sulfoxide. Biopolymers6:793-807.

29. Weinberg,R.A.,and S. Penman. 1970.Processingof 45 S nucleolar RNA. J. Mol. Biol.47:169-178.

620

J. VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/